64 bit processors compute 64 instructions at a time. Thus a 128 instruction count is processed in two passes, as opposed to 4 with a 32 bit processor. Thus it is 'faster'. This frees up resources, thus more can be spent on graphics processing (which is still done partially by CPU) and map volume information.
Completely and utterly wrong.
64bit cpu's excute instructions 64bits long as opposed to 32bits long. (When running in native 64bit mode). Their register widths are 64bit, as opposed to 32bit, so maths on very big integer numbers can be done more efficiently. The amount of instructions it performs at the same time is do with the pipeline depth, which has very little basis on the size of instructions it executes.
Also the A64 has alot more general purpose registers, so an .exe specifically compiled for it, can potentially create better code in areas that are very variable heavy. Instead of storing variables on the stack, it can keep more data in registers, which is faster.